The Composition, Structure and Origin of Proteose-peptone Component 5 of Bovine Milk

Transcription

1 Eur. J. Biochem. 9, (1978) The Composition, Structure and Origin of Proteose-peptone Component 5 of Bovine Milk Anthony T. ANDREWS Chemistry Department, National Institute for Research in Dairying, Shinfield, Reading (Received March 28, 1978) Proteose-peptone component 5 has been isolated from bovine milk. Molecular weight values within the range were obtained by sedimentation equilibrium, dodecylsulphate/ polyacrylamide gel electrophoresis and gel filtration in urea-containing buffers. A dansylation procedure showed that the sequence Arg-Glu occupied the N-terminal position while hydrazinolysis revealed C-terminal lysine. The latter was confirmed by experiments with carboxypeptidases B and C which indicated that a mixture of molecules was present, about 8% of which had a C-terminal sequence -(Ala-Met)-Ala-Pro-Lys while about 2 % had an additional -His-Lys in the terminal position. These results, together with data on the overall composition, showed that this component of the proteose-peptone fraction of milk corresponded to a mixture of molecules representing residues 1-15 and 1-17 of the B-casein molecule, a finding that was confirmed by peptide mapping. This demonstration that proteose-peptone components correspond to the N-terminal portions of the j-casein molecule while the y-caseins represent the matching C-terminal portions provides strong evidence in favour of a proteolytic mechanism for the formation of these substances iia vivo and in vitro. In a number of papers on the composition and properties of the y, R, S and TS caseins (now known as yl, yz and y3 caseins) Groves and co-workers [1,2] established that these minor casein components of bovine milk occur in a number of genetic variants and are closely related to parts of the 8-casein molecule. Subsequent work [3-61 has clearly shown that these components correspond to the C-terminal regions of the 8-casein molecule, the complete amino acid sequence of which is now known [7]. The composition and structure of these minor caseins have recently been reviewed [8] and their nomenclature clarified. Thus yl-casein (previously known as y-casein) corresponds to residues inclusive of the p-casein molecule, and there are a number of genetic variants (A, A, A3 and B) related to the corresponding variant of 8-casein in the same milk. Likewise the y2-caseins (previously termed the S and TS-A caseins) represent residues of the p-casein molecule, while the y3-caseins (formerly the R and TS-B caseins) correspond to residues At the present time the origin of these minor casein components is not known although their homologies with segments of the p-casein polypeptide chain has Enzymes. Plasmin (EC ); carboxypeptidase B (EC ); carboxypeptidase C (EC ). led to the suggestion [6,8] that they may be formed by limited proteolysis of p-casein. Indeed Eigel [9] has recently demonstrated that fragments with approximately the correct molecular weights and similar electrophoretic mobilities to yl-a2, y2-a2 and y3-a are formed by proteolysis of p-casein A in vitro with bovine plasmin. However, there is no clear evidence that in vivo proteolysis is the mechanism for formation of these caseins and if it occurs whether the proteolysis takes place within the mammary gland itself or after milking [8]. Thus de novo synthesis of these y-caseins cannot yet be ruled out, especially since the N-terminal phosphopeptide segments of the p-casein molecule which would also be expected to arise from proteolysis have never been isolated or identified in either milk or mammary gland tissues. In the course of studies on proteolysis in milk we have found that these N-terminal segments are present and are located within the proteose-peptone fraction, hence supporting a proteolytic mechanism for the formation of both the y-caseins and some of the components of proteose-peptone. The identification of proteose-peptone component 5 with residues 1-15 and 1-17 of the p-casein molecule, corresponding to the formation of the yz- and y3-caseins as the C-terminal segments, is the object of this report.

2 6 Proteose-peptone Component 5 MATERIALS AND METHODS Preparation of Proteose-peptone Component 5 Bulk fresh raw skimmed bovine milk (5 1) was heated at 95 "C for 3 min and cooled to about 3 "C. The ph was adjusted to 4.6 with 1 M HCl and precip itated casein and denatured whey proteins rem?\ LYI by centrifugation at about 15 x g for 3 min. Protein was precipitated from the supernatant liquid (3.8 1) by addition of trichloroacetic acid to 12 % and collected by centrifugation at 15 x g for 3 min. The precipitate was washed by suspension and recentrifugation once with 12 % trichloroacetic acid (2 ml) and then three times with acetone (2 ml) and air-dried to yield approximately 6.4 g of crude proteose-peptone. Gel filtration was carried out on Sephadex G-75 columns (22 x 56 mm) with.2 M sodium phosphate buffer ph 6.5 and samples of about 2 mg were applied in a volume of 3 ml. Flow rates were 8-12 ml/h and 3-ml fractions were collected. Absorbance was measured at 23nm and 28nm and the progress of the separations also monitored by examining 5-pl or 1-pl portions of fractions by polyacrylamide gel electrophoresis [lo]. Fractions rich in component 5 were pooled and trichloroacetic acid added to 1 %. The precipitated protein was redissolved in 3 ml of.2 M sodium phosphate buffer ph 6.5 containing 8 M urea and gel filtration repeated on an identical column to that used above but with buffers containing 8 M urea. Component 5 fractions were pooled, precipitated with 1 % trichloroacetic acid and washed once with 1% trichloroacetic acid and then three times with acetone as described above and air-dried. Analytical Methods Molecular weight values were obtained by gel filtration on 22 x 56-mm columns of Sephadex G-5 or Biogel P-3 in various buffers (see Results) or on Sephadex G-75 as described above. Bovine serum albumin, ovalbumin, chymotrypsinogen, b-lactoglobulin, a-lactalbumin and cytochrome c were used for column calibrations. These proteins were also used for calibration of gels for measurement of the molecular weight of component 5 by polyacrylamide gel electrophoresis in the presence of dodecylsulphate [ll]. The low-speed equilibrium technique [12] was used for measurement of molecular weight in the ultracentrifuge and the partial specific volume (212) was calculated from the composition [13,14]. For amino acid analysis, samples were heated under N2 for 24, 48 and 72 h at 11 "C in 6 M HC1 containing.1 % thioglycollic acid. Phosphate content was determined by the method of Chen et al. [15] and carbohydrate was measured by the anthrone procedure [ Dansylation of component 5 was coupled with Edman degradation to determine the N-terminal sequence [17]. Dansyl amino acid derivatives were identified by thin-layer chromatography on silica gel plates [18]. Hydrazinolysis was used for identification of the C-terminal amino acid [19]. For carboxypeptidase B treatment, enzyme (1 85 units/mg, Sigma Chemical Co.) was mixed with component 5 in.2 M N-ethylmorpholine acetate buffer, ph 8. at a molar substrate/enzyme ratio of approximately 5/1 and incubated at 37 "C. Carboxypeptidase C contained 1 unit/ml of enzyme activity (Boehringer Mannheim) and was used in.8 M sodium citrate buffer ph 5.3 at a level of.5 unit for 8.8 mg of component 5 in a volume of 5 ml. Carboxypeptidase C treatment was also applied to carboxypeptidase-b-treated component 5. After a 3-h incubation with carboxypeptidase B the ph was adjusted to 5.3 by addition of solid citric acid and.5 unit of carboxypeptidase C in.5 ml HzO added for further digestion. For peptide mapping, b-casein (3 mg) and component 5 (7 mg) were dissolved in.3 ml and.7 ml respectively of.1 M pyridine at ph ; 1 pl and 5 p1 respectively of a solution of trypsin (treated with L-l-tosylamido-2-phenylethyl chloromethyl ketone, Worthington Biochemical Corp., 3 mg/ml in.5 mm HC1) was added to give substrate/enzyme ratio of 1/1. After 24 h at 18 "C, 1-p1 portions were applied to 2 x 2-cm cellulose thin-layer plates (Merck, Darmstadt) for electrophoresis at 45 V/cm for 3 min in pyridine acetate buffer ph 3.6. Chromatography with butan-l-ol/acetic acid/h2o (4/1/5, by vol.) was used for the second dimension and peptides detected with ninhydrin. RESULTS Isolation of Component 5 Gel filtration of the crude proteose-peptone mixture is shown in Fig. 1 while a diagram of the polyacrylamide gel electrophoresis patterns of various fractions is shown in Fig. 2. On the basis of these patterns, gel filtration fractions were pooled as indicated (A- F on Fig. 1) and protein precipitated by adding trichloroacetic acid to 12.5%. Yields of acetone-dried material from 26 mg of crude proteose-peptone mixture were : pool A 87.7 mg, B 24. mg, C 48.5 mg, D 4. mg, E 12.7 mg and F 15.5 mg. It was estimated that component 5 comprised between 19 %and 25 %of the total proteose-peptone fraction. Pool B material was then further purified by gel filtration in the presence of urea (Fig. 3) and fractions containing component 5 pooled, as indicated by the bar, dialysed overnight

3 A. T. Andrews 61 A B C D E F H I : : : : H 2.5 E c 4 2D N k I M e o 1.5 R p" 1..5 O Fraction number Fig. 1. Gel,fifrsatiorr of LI twde [?r~i"tt'c~s~-~?~~~tnnr mi-rturc oil a dunin (22 x 56 nim) of Sephadex G-75 in.2 M sodium phosphate buffer pli 6.5. Fractions of 3 ml collected at a flow rate of about 1 ml/h. Fractions were pooled as indicated by the bars A - F. Absorbance measured at 28 nm after application of 26 mg proteose-peptone to the column (-); absorbance at 23 nm with a column load of 31 mg (----) ' Fraction number Fig. 3. Gel,f2[rufion of' poi B imrtv.irif nf Fig. 1 on a column (22 x 56 mni) fsephade.x G-75 in.2 M.sodium phosphare bufler ph 6.5 containing 8 M urea. Flow rate was 1 ml/h and 3-ml liactions were collected. Fractions containing component 5 were pooled as indicated by the bar Origin - Component 5 - Component 8 F - I Mix Fraction numbet I Fig. 2. Diagrammatic representation of' polyacrylamide gel e1ectrophorc~si.s patterns obtained from various fractions eluted,from the Sephadex G-75 column shown in Fig. 1. Electrophoresis was performed in 9.4y4 gels (T = 9.4%; C = 4.25%) at ph 8.9 [lo]. Mix = unfractionated proteose-peptone mixture versus HzO (using.25-in-diameter Visking tubing) and the protein precipitated by addition of trichloroacetic acid to 12.5%. After washing with trichloroacetic acid and acetone the yield of component 5 was 14.5 mg. Molecular Weight Gel filtration on a Sephadex G-5 column with a buffer of.2 M NH4HC3 ph 8. indicated a molecular weight of 275 zk 15. A similar value (31 k 3) was also obtained in.2 M NH~HCOJ buffer ph 8. containing.1 M EDTA on a Biogel P-3 column. While gel filtration at room temperature consistently gave these values in aqueous buffers, other methods of estimating molecular weight led to quite different answers. Polyacrylamide gel electrophoresis in the presence of dodecylsulphate gave 13 f 2, which agreed well with the value of 135 k 1 given by gel filtration on Sephadex G-75 in the presence of 8 M urea and.2 M 2-mercaptoethanol. By sedimentation equilibrium at 2 "C in an aqueous buffer (.2 M sodium phosphate ph 6.5) component

4 62 Proteose-peptone Component 5 Table 1. The composition of proteose-peptone component 5 Samples of component 5 (approximately.5 pmol) were hydrolysed under Nz in 1.O ml 6 M HCI containing.68 4 thioglycollic acid for 24, 48 and 72 h at 11 "C. Recoveries were within the range 87-96%. Figures in parentheses in the last column are those obtained when amino acid residues 16 and 17 are included. Values for threonine and serine are extrapolated values. Phosphate was determined colorimetrically [I r Amino acid Amount Nearest,/&Casein A' integer residues 1-15 (17) mol/l2 3 g Aspartic acid Threonine Serine Glutamic acid Proline Glycine Alanine Valine Methionine Isoleucine 6.66 I I Leu c i n e Tyrosine Phenylalanine Histidine (2) Lysine (8) Arginine Fig.4. Release of C-terminal amino acids from component 5 by treatment with carboxypeptidase B at ph 8.. Liberation of lysine () and histidine () Phosphate had a molecular weight of 119 k 5 when a calculated value of.726 ml/g was used for the partial specific volume (uzj based on the composition shown in Table 1. Composition of Component 5 The amino acid composition (corrected for 9.5% moisture content) is shown in Table 1 from which it may be seen that component 5 contained large amounts of glutamic acid and proline while tryptophan, cysteine and cystine were absent. Tyrosine content was also low so that the molar absorption coefficient at 28 nm was only 33 M-' cm-'. The absorption maximum was at 275 nm (E = 3432 M-' cm-'j, there was a minimum at 255 nm (E = 2145 M-' cm-'j and the ratio of absorption at 28 nm to that at 26 nm was Component 5 contained 1.24 by weight of phosphorus, equivalent to about 5 mol/mol (Table 1 j. This phosphate was alkali-labile and completely released by treatment with.2m NaOH at 1 "C for 2 h, suggesting that it probably occurred linked to serine or threonine. Only trace amounts of carbohydrate (less than.8 mol/molj were found by the anthrone method and no glucosamine or galactosamine could be detected with the amino acid analyser. Time (h) Fig. 5. Time course measurements showing the liberation of a number of amino acids from the C-terminal end of component 5 by the action of carboxypeptidase C at ph 5.3. Release of lysine ( k O ), proline (*---O), alanine (O----O), methionine (-.), histidine ( x - x ), glutamic acid (A----A), valine (A----A) and serine (+ -~+) Terminal Amino Acid Analyses and Peptide Maps Arginine occupied the N-terminal position, the dansyl-arginine derivative being obtained in good yield with no accompanying traces of any other amino acid derivatives. After one cycle of the Edman degradation and dansylation of the product, dansyl-glutamic acid was the only amino acid derivative identifiable but yields were small. After a further cycle, no dansylderivative could be identified, only small traces of

5 A. T. Andrews A Q O I 1 -Electrophoresis -Electrophoresis Fig. 6. Cellulose thin-layer peptide maps of trypsin digests of (A) component 5 and (B) p-casein. Electrophoresis was carried out in 1 M acetic acid adjusted to ph 3.6 with pyridine, run at 45 V/cm for 3 min, in the horizontal axis and with chromatography in the vertical dimension using the organic phase of butan-l-ol/acetic acid/h2 (4/1/5, by vol.). (A) Map obtained from 1 ~1 digest containing initially 93 pg of component 5, (B) 1 p1 /3-casein digest, corresponding to 97 pg of /3-casein + fluorescent impurities or by-products forming streaks being seen on the thin-layer chromatography plates. Thus these results suggested N-terminal arginine linked to glutamic acid. Hydrazinolysis released lysine as the only free amino acid from the C-terminal position, although the yield (.17 mol/mol) was not particularly good. C- terminal lysine was confirmed by carboxypeptidase B (Fig. 4) which released approximately.25 mol/mol within 3 min,.44 mol/mol in 1 h and a maximum of 1.18 mol/mol after 4 h. Carboxypeptidase B also released some histidine, amounting to.13 mol/mol in 3 min,.1 8 mol/mol after 1 h and reaching the maximum (.21 mol/mol) in less than 2 h. Thus after 3 min 62% of the maximum amount of histidine was released while only 21 % of the final concentration of lysine had been freed. After 1 h these figures were 86 and 37 % respectively, so that while total amounts of lysine released were greater throughout the incubation period, histidine reached its maximum value more quickly. Taken together with the hydrazinolysis experiments these results could best be explained by suggesting that, as currently prepared, component 5 was a mixture in which about 8% of the molecules end with -X-Lys and 2 o/, with -X-Lys-His-Lys. In a further experiment, treatment of a sample of component 5 with carboxypeptidase B for 3 h liberated lysine and histidine and removed this heterogeneity giving a population of molecules ending only at -X. This product was then treated with carboxypeptidase C which rapidly liberated proline to the extent of.31 mol/mol in 3 min and about.5 mol/mol after a 2-h incubation. Alanine was released more slowly (.9 and.28 mol/mol after 3 min and 2 h respectively) with small quantities of methionine (very approximately.13 mol/mol in 2 h) also being formed. Carboxypeptidase C treatment of intact component 5 is shown in Fig.5. It can be seen that, as expected, there was an initial rapid release of lysine. Proline and alanine were also rapidly released, initially proline being freed more rapidly than alanine but after 2 h or more the amounts of alanine exceeded those of proline. Methionine was released more slowly than any of these residues, but still in amounts exceeding.5 mol/mol after a 4-h incubation. These findings are best explained by the C-terminal sequence -(Ala,Met)-Ala-Pro-Lys. While further interpretation then becomes somewhat ambiguous, it is interesting to note that the much slower release of glutamic acid, valine and serine was entirely in agreement with the postulated sequence -Ser-Lys-Val-Lys-Glu-Ala-Met- Ala-Pro-Lys, corresponding to residues of /3-casein. The release of histidine was again consistent with a mixture of molecules, a small proportion of which included residues 16 and 17 of,&casein. Twodimensional maps of the tryptic peptides from component 5 and from p-casein are shown in Fig.6. All the peptides from component 5 corresponded to peptides which were also formed from,%casein, although in the latter case the total number of peptides was

6 64 Proteose-peptone Component Arg-Glu-Leu-Glu-Glu-Leu-Asn-Val-Pro-Gly-Glu-Ile-Val-Glu-SerP-Leu-SerP-SerP-SerP Glu Glu-Ser-Ile-Thr-Arg-Ile-Asn-Lys-Lys-Ile-Glu-Lys-Phe-Gln-SerP-Glu-Glu-Gln-Gln-Gln Thr-Glu-Asp-Glu-Leu-Gln-Asp-Lys-Ile-His-Pro-Phe-Ala-Gln-Thr-Gln-Ser-Leu-Val-Tyr Pro-Phe-Pro-Gly-Pro-Ile-~-Asn-Ser-Leu-Pro-Gln-Asn-Ile-Pro-Pro-Leu-Thr-Gln-Thr Pro-Val-Val-Val-Pro-Pro-Phe-Leu-Gln-Pro-Glu-Val-Met-Gly-Val-Ser-Lys-Val-Lys-Glu- 15 i16 17 t Ala-Met-Ala-Pro-Lys-~-Lys-Glu-Met-Pro-Phe-Pro-Lys-Tyr-Pro-Val-Gln-Pro-Phe-Thr Glu-Ser-Gln-Ser-Leu-Thr-Leu-Thr-Asp-Val-Glu-Asn-Leu-His-Leu-Pro-Pro-Leu-Leu-Leu Gln-Ser-rrp-Met-His-Gln-Pro-His-Gln-Pro-Leu-Pro-Pro-Thr-Val-Met-Phe-Pro-Pro-Gln Ser-Val-Leu-Ser-Leu-Ser-Gln-Ser-Lys-Val-Leu-Pro-Val-Pro-Glu-Lys-Ala-Val-Pro-Tyr Pro-Gln-Arg-Asp-Met-Pro-11e-Gln-Ala-Phe-Leu-Leu-Tyr-Gln-Gln-Gln-Pro-Val-Leu-Gly-Pro- 29 -Val-Arg-Gly-Fro-Phe-Pro-Ile-Ile-Val. Fig.7. Amino acid sequence of'bovine 8-casein A' showing (UYYOM'S) points of cleavage resulting in the formation of yl, yz and y3 caseins [7]. Residues underlined represent the sites of amino acid substitutions or deletions in known genetic variants about twice as great. This was confirmed by running samples of component 5 digest and fi-casein digest together on the same plate, which showed conclusively that component 5 represents part of the B-casein molecule. DISCUSSION The primary structure of fi-casein A2 is shown in Fig. 7, and from this it can be seen that the analytical data presented here is entirely consistent with the identification of component 5 with residues 1-15 and 1-17 of the fi-casein sequence. The overall amino acid composition and phosphate content (Table 1) is almost precisely in agreement with that of the N- terminal half of P-casein. The serine content is a little low probably because five of these residues occur as phosphoserine and in our experience recoveries of serine from phosphoserine are not as high as when it occurs unphosphorylated. The small discrepancy in proline is probably explained by the likely occurrence of genetic variants (see below) which will result in a lower content of proline and increased histidine with respect to the values for fi-casein A2. Valine and isoleucine tend to be more resistant to acid hydrolysis than most other amino acids and inspection of Fig. 7 reveals the sequences -1le-Val- at position 12 and 13 and -Val-Val-Val- at position which are likely to be especially resistant to hydrolysis, hence accounting for the slightly low recoveries of these amino acids. The amino acid composition is broadly similar to that reported by Kolar and Brunner [2] but our preparation of component 5 was clearly not a glycoprotein. These workers [2] also suggested a molecular weight of 14 3 obtained from sedimentation equilibrium analysis in fair agreement with our value of The calculated value for residues 1-15 of /?-casein A2 is 12258, which is very similar to the values we obtained in the ultracentrifuge, by dodecylsulphate/ polyacrylamide gel electrophoresis or by gel filtration using urea-containing buffers. In aqueous buffers without urea, however, much higher apparent molecular weight values were obtained by gel filtration. These were not influenced by the properties of the gel matrix (Sephadex and Biogel gave similar results), by high ionic strength (1 M NaCl), by ph over the range or by addition of.2 M EDTA to remove traces of calcium or other heavy metal ions. One explanation for this could be that component 5 exists in a dimeric form, but this was not consistent with the ultracentriftige results and any monomerdimer equilibrium should be apparent in both systems. Another explanation, which we favour at present, is based on conformational considerations. The caseins have high proline contents and are often regarded as proteins with little or no tertiary structure which exist in a largely random coil conformation. Thus since component 5 represents the N-terminal half of the /?-casein molecule it too may be expected to exist largely as a random coil type of structure. In aqueous buffers it is conventional, and usually correct, to cali-

7 A. T. Andrews 65 brate gel filtration columns for molecular weight estimation with globular proteins of known size. When the unknown protein exists as a random coil such calibration is clearly inappropriate, but when dissociating buffers are used the standard proteins will also have random coil conformations and the correct molecular weight should be obtained. The component 5 preparation used in this work migrated as a single, rather broad, band during polyacrylamide gel electrophoresis. There were two probable causes for the broadness of the band. Firstly, the milk was heated in the first stage of the preparation and it is possible that this may have caused some loss of the amide groupings of asparagine and glutamine and possibly some loss of phosphate from the phosphoserine residues. Secondly, the presence of closely related molecules of different primary sequence. As discussed in Results, it is probable that our material consisted of a mixture of molecules representing residues 1-15 and 1-17 of,&casein, but in addition to this there should be some heterogeneity due to the different genetic variants. Proteose-peptone samples were prepared from bulk milks from a herd of cows which were mostly heterozygous for p-casein and therefore contained two of the three variants, A', A2 and B, with a few animals homozygous for one of these variants. Only two animals out of 8 gave milk containing b-casein A3 and no other genetic variants were present at all. The only genetic differences to be expected in component 5 samples therefore would be due to substitution of histidine for proline at position 67 for both the A' and B variants but otherwise there are no differences in sequence from that shown in Fig. 7 within the region 1-15 and 1-17 (the very small amount of,&casein A3 which involves a substitution of glutamine for histidine at position 16 can probably be ignored). Preliminary experiments using longer (1 1-cm) polyacrylamide gels have shown that component 5 can be resolved into at least three closely spaced bands, although further work is required to establish the basis for this heterogeneity. The demonstration that component 5 and proteosepeptone component 8F [21] represent the N-terminal portions of the fi-casein molecule while yl, y2 and y3- casein represent the corresponding C-terminal portions provides very strong evidence that both these proteose-peptones and the y-caseins are formed by a proteolytic breakdown mechanism from p-casein. The author thanks Mr M. D. Taylor for skilled technical assistance, Mrs D. J. Knight for performing the amino acid analysis and Mrs V. A. Hill for the sedimentation equilibrium runs. REFERENCES 1. Groves, M. L. & Gordon, W. G. (1969) Biochim. Biophys. Acta, 194, Groves, M. L. & Kiddy, C. A. (1968) Arch. Biochem. Biuphys. 126, 1% Gordon, W. G., Groves, M. L., Greenberg, R., Jones, S. B., Kalan, E. B., Peterson, R. F. & Townend, R. E. (1972) J. Dairy Sci. 55, Groves, M. L., Gordon, W. G., Kalan, E. B. & Jones, S. B. (1972) J. Dairy Sci. 55, Groves, M. L., Gordon, W. G., Kalan, E. B. & Jones, S. B. (1973) J. Dairy Sci. 56, Gordon, W. G. & Groves, M. L. (1975) J. Dairy Sci. 58, Ribadeau-Dumas, B., Brignon, G., Grosclaude, F. & Mercier, J.-C. (1972) Eur. J. Biochem. 25, Whitney, R. McL., Brunner, J. R., Ebner, K. E., Farrell, H. M., Josephson, R. V., Morr, C. V. & Swaisgood, H. E. (1976) J. Dairy Sci. 59, Eigel, W. N. (1977) Int. J. Biochem. 8, Hillier, R. M. (1976) J. Dairy Res. 43, Weber,K. & Osborn,M. (1969) J. Biol. Chem. 244, Van Holde, K. E. (1967) Fractions, no. 1, Beckman Instruments Inc., Palo Alto. 13. McMeekin, T. L., Groves, M. L. & Hipp, N. J. (1949) J. Am. Chem. Soc. 71, Charlwood, P. A. (1957) J. Am. Chem. Soc. 79, Chen, P. S., Toribara, T. Y. & Warner, H. (1956) And. Chem. 28, Yemm, E. W. & Willis, A. J. (1954) Biochem. J. 57, Gray, W. R. (1967) Methods Enzymol. 11, Seiler, N. (197) Methods Biochem. Anal. 18, Fraenkel-Conrat, H. & Tsung, C. M. (1967) Methods Enzymol. 11, Kolar, C. W. & Brunner, J. R. (197) J. Dairy Sci. 53, Andrews, A. T. (1978) Eur. 1. Biochem. 9, A. T. Andrews, National Institute for Research in Dairying, Shinfield, Reading, Great Britain, RG2 9AT